Emotion Recognition for User Centred E-Learning
Oryina Kingsley Akputu, Abiodun O. Adedolapo · 2016
A Vision of most e-learning models is to accurately recognize learner's post (pre) learning feedbacks to improve learning interaction. Several effort towards user centred e-learning have been made in literature, but mostly concentrates on cognitive based feedbacks for learner's modelling. However, Beside cognitive factors, emotions of the learner are equally important but seldom neglected. This paper present a new architecture of FERT. A processing pipeline of the FERT is realized through results of a preliminary analysis across selected facial features descriptive techniques, namely, Gabor wavelet, Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA). Further more, Emotion recognition of the FERT scheme employs Multiple Kernel Learning (MKL) framework which reportedly outperforms traditional classifiers. Experiments have been conducted on contextual emotion datasets and results shows good performances of the FERT scheme. Finally, a conceptually user centred e-learning model that has potentials for improving learning interaction is described.